What a small local AI model taught me about context
Running a small AI model on my own hardware was a reminder to give AI less context and clearer steps. What that means for nonprofit teams.
Running a small AI model on my own hardware was a reminder to give AI less context and clearer steps. What that means for nonprofit teams.
How AI agents can carry the admin load through GivingTuesday and year-end giving, and the limits to set before they touch tax receipts.
Impact data is messy and scattered. How AI can turn case notes into an impact record your board can trust, with people still signing off.
AI has no common sense about your organization. How to give it clear instructions with a four-part prompt framework, and a safe place to work.
Treating AI like a human colleague leads to disappointment. A better mental model, where AI struggles and does well, and a safe first exercise.
Efficiency comes down to small daily tasks. What AI connected to your CRM, drives and email can take off fundraisers, program staff and marketers.
A prompt isn’t a guardrail. What real AI guardrails look like, and why they help a nonprofit team trust AI with real work.
How to let AI help with your CRM and donation systems safely, by sorting each change into a risk tier that sets how much the AI can do alone.